1) Field of the Invention
The present invention relates to a technology for setting image-capturing conditions to efficiently recognize identification data of a moving object.
2) Description of the Related Art
Traffic surveillance system to read out registration information written on a license plate of a vehicle is currently used to monitor traffic of vehicles. In the traffic surveillance system, an image of the license plate is captured, and a character recognition process is performed to the registration information on the image captured. The result of the character recognition process with place and time information is stored in a database. Thus, information regarding a specific vehicle's traveling can be easily retrieved from the database.
As an example of the traffic surveillance system, a mobile character recognition system that is portable in a vehicle and setup at an arbitrary place is disclosed in Japanese Patent Laid-Open Publication No. 2003-22494. When searching for a particular vehicle, the mobile character recognition system first checks whether any existing registration information input earlier matches with character-recognized registration information. If matching registration information is found, the mobile character recognition system displays corresponding information.
However, with the conventional technology, it is not possible to adjust conditions such as shooting timing when a road is assigned to one-way road according to time or traffic condition, or when traffic is allowed to flow in the opposite direction on a part of a multi-lane road, resulting in inefficient character recognition of the license plate.
For instance, when a vehicle is approaching a camera, it is desirable to wait until the vehicle gets to a position where the license plate on the front of the vehicle is readable before capturing an image. Likewise, when the vehicle is passing from behind the camera, it is desirable to wait until the vehicle moves away to a position where the license plate on the back of the vehicle is readable before taking the image.